Web Based Evaluation, Design and Updating of a Reservoir Control System

Abstract

Optimal hydrocarbon production is facilitated by a continual adjustment of wellbore pressure or flow rate, most easily implemented with an automated control system. For each well, this system consists of a tuned logic controller, a remotely operable throttle valve, and sensors such as pressure transducers and/or flow meters. Joint modeling of the reservoir and the control components is necessary for tuning the feedback algorithm in order to have a stable and acceptable rate of response (SPE 84219). Given the target audience, a platform independent user-friendly front-end is desirable; but the system evaluation is computationally intensive, thus necessitating a sophisticated back-end server. To this end, we present a multi-tier Web application that offers a universal interface to a simulator running on a central compute server. The rendering of the Web interface is driven by the data. The application delivers the simulation input data to the compute server through a relational database. The results may be examined online and parameters adjusted until a satisfactory step response is obtained. The method is also applicable for downloading tuned parameters to a semi-autonomous wellsite controller. Introduction Smart completion systems may be operated by continuous or periodic throttling of downhole valve assemblies.1,2 By adjusting the resistance to flow from various formation intervals, production may be "controlled." Various objectives may be desired, but the primary one is likely to be the reduction of water to oil ratio (WOR). A decision to install a smart completion to achieve this objective will be dictated by economics - a cost/benefit issue. Oil production in relation to water may be enhanced by altering sweep, i.e., adjusting injection or production rate ratios of patterned wells, or by changing vertical contact of the invading fluid by throttling sections within a well. It has been argued that the former is principally a flow control requirement, whereas the latter is exercised primarily as a pressure control mechanism.3 Regardless of which mechanism is implemented, it is expected that an optimal production plan4,5 would dictate a flow rate or a pressure for injection and production at the smart wells. A system to implement either of the control mechanisms requires appropriate sensor measurements, which upon comparison with the desired value will generate a deviation or an error. Both the instantaneous and the historical values of the error signal may be used by a control algorithm to drive a smart completion such that the error is reduced to arbitrarily small values. Controllers consisting of a Proportional, Integral and Derivative (PID) components, or their digital versions, function in the above mode.6 A block diagram illustrating the feedback system is given in Fig. 1. A significant aspect of controller analysis is the design of a stable but responsive controller algorithm. For example, in a PID system, the magnitude of the constants driving each of the components is "tuned" to achieve an acceptable response. Various techniques for carrying this out are well known,6 a number of which are applicable to linear, drift-free systems.